Marshall JCM2000 DSL100 Bias Drift Mod - Didn't fix bias drift!

[bw1985]

Hi all,

Was hoping to get some insight into what might be going on with this Marshall JCM2000 DSL100 I am working on. I took the necessary steps to remove the grid resistors from the board, cut the traces and mount new 5.6k resistors upright on pin 5 of each power tube and ran the wires above the board. At first, it seemed like the bias drift problem was alleviated, but slowly the bias started to creep up and up. On one set of tubes it settled around 90mV after an hour or so of running hot (bias set at lowest point possible). On another set of tubes, the left side pair went up to about 130mV. I am unsure if there is an issue with that set of tubes, though they technically "pass" in my basic tester.

The board rev on this one is the original i think (JCM2-60-00). The issue appears to be "better", but I am not sure if all of the tubes I have are bad, or if some conductance is happening somewhere else now. If anyone has experienced this after performing the mods, let me know and if there is a solution.

Thanks and all the best,
BW

[Stratfordade]

Monitoring the grid voltage should tell you if the bias is stable, or if there is some other runaway behaviour happening with you particular EL34s. Also swap over the output pairs and see if it rises to 130mV on the right side too.

[bw1985]

Monitoring the grid voltage should tell you if the bias is stable, or if there is some other runaway behaviour happening with you particular EL34s. Also swap over the output pairs and see if it rises to 130mV on the right side too.

The grid voltage seems quite stable and even across all 4 tubes. It could be the case of the tubes themselves. I just find it odd that I can't bias the tubes lower than 85-90mV.

[Jon Snell]

To save time and speed things up, I normally check for bad connections on the bias board, (pots not soldered etc), then replace the grid coupling capacitors C6 & C7. (22n 630v).
Then watch what happens. The G1 voltage on pins 2 & 6 of the bias board, must remain stable with no signal present after setting the idle current you desire.

[bw1985]

To save time and speed things up, I normally check for bad connections on the bias board, (pots not soldered etc), then replace the grid coupling capacitors C6 & C7. (22n 630v).
Then watch what happens. The G1 voltage on pins 2 & 6 of the bias board, must remain stable with no signal present after setting the idle current you desire.

Thank you for that advice. I may end up replacing them as a precaution. I did however just swap tubes with another set (a known good set) and now my starting bias is reading around 55mV per side. Perhaps I just have a handful of bad tubes. I know at least a couple of them redplated and cooked when the amp blew fuses.

[Mick Bailey]

See how it goes, but in some cases the board can leak voltage between pins. I machined up a bit - just a piece of maybe 3/16" tool steel drilled out clearance for the socket pin and some serrated teeth filed into the end (sawtooth shape). Then hardened right out on the end. In a drill this quickly cuts a neat hole around pin 5 and leaves it isolated ready to connect the resistor.

You could carefully do the same sort if job with a Dremel and suitable bit.

This is a last resort where everything else has been done but the problem remains. It has saved quite a few boards over the years.

[bw1985]

See how it goes, but in some cases the board can leak voltage between pins. I machined up a bit - just a piece of maybe 3/16" tool steel drilled out clearance for the socket pin and some serrated teeth filed into the end (sawtooth shape). Then hardened right out on the end. In a drill this quickly cuts a neat hole around pin 5 and leaves it isolated ready to connect the resistor.

You could carefully do the same sort if job with a Dremel and suitable bit.

This is a last resort where everything else has been done but the problem remains. It has saved quite a few boards over the years.

I was concerned about leakage between pins as well. So far I've run my known set of "good" tubes for 2 hours and bias stayed stable around 75mV on both sides after adjusting at the 20/30min mark. I think I may be good to go. I didn't actually drill out pin 5 of each board, but will do so as a last resort.

[J M Fahey]

Sad to say that if you have one of the dreaded conductive boards, you may try all of the "home medicines" suggested in various Forums and Videos but only true solution is full board replacement.

Situation was so bad that Marshall themselves offered replacement boards.

[Stratfordade]

I was concerned about leakage between pins as well. So far I've run my known set of "good" tubes for 2 hours and bias stayed stable around 75mV on both sides after adjusting at the 20/30min mark. I think I may be good to go. I didn't actually drill out pin 5 of each board, but will do so as a last resort.

if bias is stable at that level after two hours all sounds good to me, and that some of the valves in the ‘bad’ set were damaged in previous bias runaway.